[U-Boot] mvebu: reserve SRAM memory on Marvell Armada 3700/7K/8K

[Heinrich Schuchardt]

On 6/13/19 7:48 AM, Stefan Roese wrote:
> Added Heinrich to Cc (and use Alex's new address)
>
> On 11.06.19 13:00, Patrick Wildt wrote:
>> The ARM-TF and the optional OP-TEE use the memory region 0x4000000
>> to 0x5400000 and should be reserved in the memory map, otherwise the
>> OS might wrongly assume that it can use that memory area for itself.
>> This has also been done in EDK2 [0].
>>
>> [0]
>> https://github.com/tianocore/edk2-platforms/commit/bf1c4a2cf8024669d1748e78c7e417433f85707e
>>
>>
>> Signed-off-by: Patrick Wildt <patrick at blueri.se>
>>
>> diff --git a/arch/arm/mach-mvebu/arm64-common.c
>> b/arch/arm/mach-mvebu/arm64-common.c
>> index aaf7b7c447..7572aad8c9 100644
>> --- a/arch/arm/mach-mvebu/arm64-common.c
>> +++ b/arch/arm/mach-mvebu/arm64-common.c
>> @@ -14,6 +14,7 @@
>>   #include <asm/arch/cpu.h>
>>   #include <asm/arch/soc.h>
>>   #include <asm/armv8/mmu.h>
>> +#include <efi_loader.h>
>>     DECLARE_GLOBAL_DATA_PTR;
>>   @@ -142,5 +143,11 @@ int arch_early_init_r(void)
>>       pci_init();
>>   #endif
>>   +#ifdef CONFIG_EFI_LOADER
>> +    /* Reserve trusted SRAM section */
>> +    efi_add_memory_map(0x04000000, 0x01400000 >> EFI_PAGE_SHIFT,
>> +               EFI_RESERVED_MEMORY_TYPE, false);
>> +#endif

We already have a reservation via the device tree for a part of this
area [0x4000000-0x4200000[ in U-Boot and Linux:

cf63dad014bae080445bccbf9cecbe05f2cbed45
arm64: dts: marvell: armada-ap806: reserve PSCI area

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v5.2-rc4&id=132ac39cffbcfed80ada38ef0fc6d34d95da7be6

As the concerned boards can be booted not only via bootefi but also via
booti I suggest to update the device trees in Linux and U-Boot.

Best regards

Heinrich

>> +
>>       return 0;
>>   }
>>
>
> I would like to see some comments from the U-Boot "EFI guys", if
> this is the correct approach. I remember some discussions about
> using "reserved-memory" in the DT for this but I might be wrong
> here.
>
> Any comments?
>
> Thanks,
> Stefan
>